ریشه پولی اطلاعات نامتقارن در بازارهای سهام بین المللی

Existing studies using low-frequency data have found that macroeconomic shocks contribute little to international stock market covariation. However, these papers have not accounted for the presence of asymmetric information where sophisticated investors generate private information about the fundamentals that drive returns in many countries. In this paper, we use a new microstructure data set to better identify the effects of private and public information shocks about U.S. interest rates and equity returns. High-frequency private and public information shocks help forecast domestic money and equity returns over daily and weekly intervals. In addition, these shocks are components of factors that are priced in a model of the cross-section of international returns. Linking private information to U.S. macroeconomic factors is useful for many domestic and international asset-pricing tests.

The causes of international stock market covariation remain a puzzling issue in finance. Asset-pricing models dictate that expected stock returns vary in response to changes in risk-free interest rates, changes in expected future cash flows, and/or changes in the equity risk premium. In a rational asset-pricing framework, with international market integration, comovements in international stock returns would be driven by news about macroeconomic factors that affect cash flows, risk-free rates, or risk premiums in many countries. Existing studies using low-frequency data, however, show that macroeconomic factors have a limited impact on international equity returns. For example, King et al. (1994) construct a factor model of 16 national stock market monthly returns and examine the influence of 10 key macroeconomic variables. They conclude that the surprise components of these observable variables contribute little to world stock market variation. Rather, there is a dominant unobservable (i.e., non-public) factor driving international returns. 1 Karolyi and Stulz (1996) show that neither macroeconomic news announcements nor interest rate shocks significantly affect comovements between U.S. and Japanese stock returns. Connolly and Wang (2003) examine the open-to-close equity market returns of the U.S., the U.K., and Japan and find that foreign returns cause movements in domestic markets even after accounting for macroeconomic news announcements. If public news about macroeconomic variables is not responsible for the comovements, could some “market friction” be responsible? One potential friction is trading based on asymmetric information. In the literature that examines the limits to international risk-sharing, asymmetric information is used as a theoretical explanation of the “home bias” and “familiarity” puzzles in international portfolio selection.2 In contrast, the literature has been largely silent on the effects of trading based on private information on international asset-return covariability. When sophisticated agents trade, their private information is (partially) revealed to the market, causing revisions in asset prices. Trading based on private information could thus be a potential cause of the comovements in international stock returns if agents had superior knowledge about the common factors that price equities from many countries. However, the economic origins of such private information remain unexplored. Indeed, Goodhart and O'Hara (1997) wonder: “in the international context, how could private information be expected to have a global impact?” In this paper, we provide an answer to this question. We start by providing empirical estimates of trading based on private information in the U.S. money and equity markets. Using an analysis of microstructure data, we show that some agents have superior knowledge about both future U.S. interest rates and aggregate U.S. equity market returns. Trades based on private information have a significant impact on money and equity market returns over holding periods that range from one day to one week. Our interpretation of these results is that sophisticated investors have good information about future macroeconomic factors that will affect both U.S. equity prices and interest rates. If international equity markets are integrated, then information about U.S. factors will give informed agents superior knowledge about the global factors that price stocks in many countries (Albuquerque et al., 2006). It is then likely that the private information of the sophisticated investors trading in these (liquid) U.S. markets will help explain the cross-section of international equity returns. We then use a latent-factor model of international equity returns to evaluate the effects of private information originating in U.S. markets on foreign markets. The latent factor is composed of public and private information shocks from the U.S. money and equity markets. We examine how the factor is related to daily and weekly returns on foreign equity markets. We show that private and public information shocks arising in U.S. markets are components of the priced factor in a model of the cross-section of international equity returns. Sophisticated investors have an impact on global markets when their superior information is incorporated into international equity returns. To the best of our knowledge, this is the first paper to show that private information is part of a priced factor in an international setting. This paper extends the existing literature in three ways. The first extension is to adapt techniques from the microstructure literature (primarily Hasbrouck, 1991) to identify information shocks in aggregate U.S. stock and money markets from quote revisions and order flows sampled at high frequencies. These shocks can be orthogonalized into those due to private and public information. Using order flow sampled at high frequencies provides a powerful method of obtaining private information shocks (e.g., Hasbrouck, 1991, Brennan and Subrahmanyam, 1996, Madhavan et al., 1997 and Yu, 2003). Similarly, the effect of public news on asset prices can be measured more accurately using high-frequency data ( Andersen et al., 2003). We can use the high-frequency shocks obtained from our time-series regressions as elements of the latent factor in a conditional asset-pricing model of the cross-section of international returns. The model reveals that the shocks are both statistically and economically important. The second extension is to focus on shocks related to a specific macroeconomic fundamental: U.S. interest rates. Cochrane and Piazzesi (2002) and others have shown that unanticipated daily movements in short-term U.S. interest rates are good proxies for monetary policy innovations. Cochrane and Piazzesi (2002) also show that these public information shocks have a large impact on short and long-term U.S. bond yields. We extend their work in two ways. First, using time-series regressions on our microstructure data, we obtain both public and private information shocks about U.S. interest rates. We then show that these shocks affect prices in the U.S. equity market. Second, we show that the shocks have an effect on the cross-section of international equity returns. Our third extension is that, instead of focusing on the returns of foreign stocks traded in national markets, we use foreign stock indexes that trade in New York as exchange traded funds (ETFs). ETFs are bundles of foreign stocks that trade on the American Stock Exchange (AMEX) and are priced in U.S. dollars. They are designed to be a low-cost instrument that tracks a foreign stock index. Because the supply of an ETF can be altered at any time, arbitrage ensures that its price closely tracks the index. By using these contracts, we observe returns and order flows on foreign equities that trade contemporaneously with their U.S. counterparts. We can thus obtain high-frequency measures of public and private information shocks that affect the foreign indexes during U.S. trading times. One potential problem in previous low-frequency studies of international equity market comovements is non-synchronous trading. Those studies examined the impact of information events that occurred during U.S. trading times on foreign (overnight) returns. The different time zones of the markets complicate the inference.3 In this paper, we avoid this problem by using foreign assets that trade contemporaneously with American stocks, and by focusing on public and private information shocks released during U.S. trading times only.4 The foreign assets, however, will respond to news released in the home markets as well. Thus, our approach does not measure the effects of all trading based on private information on the assets, but only a subset of the trades. A potential problem with our approach is that the foreign indexes (ETFs) trade less in the U.S. market than they do in their home market. The potential for stale prices will complicate any short-run analysis of the returns. In contrast, the focus in this paper is on changes in the foreign equity prices over holding periods ranging from one day to one week.5 At these intervals, stale prices will have much less of an influence. An additional benefit of examining daily and weekly intervals is that we obtain the aggregate U.S. and foreign market response to information released during U.S. trading times. This is important, because foreign market trading could negate the effects of information released during U.S. trading times, leaving no long-run impact on the price of the security. We find, however, that our high-frequency information shocks that occur during U.S. trading times continue to have an impact on foreign equity returns up to a week later. We proceed as follows. In Section 2, we describe the data. In Section 3, we show that trading based on private information predicts both U.S. money and equity market returns over daily and weekly holding periods. In Section 4, we present our latent-factor model and show that private information is a priced factor in the cross-section of international equity returns. We conclude in Section 5.